yep, that don't look right.
No clue, though looking at the photo in your previous thread... Did you fix the battery corrosion ?

Ouch. That poor thing is rediculously slow. Have you tried to run the tests with cache disabled and perhaps even physically removed from the board?
I'd parity active or disabled?

For comparison:
I've got a Biostar board with DX40 soldered.
Cachechk: within cache: 37.5 ms/b; outside cache: 78.7 ms/b
Landmark shows 70MHz AT.
My 286-16 even shows up as 24MHz AT
My Asus 486SX with both L1 and L2 disabled shows 14MHz AT in Landmark and 88ms/b in Cachechk.

I'd look into the cache with these figures.

weedeewee wrote on 2022-12-17, 21:08:

yep, that don't look right.
No clue, though looking at the photo in your previous thread... Did you fix the battery corrosion ?

I removed the barrel battery and attached an external one. May look to see if there is a connection there to the cache.

waterbeesje wrote on 2022-12-17, 21:24:

Ouch. That poor thing is rediculously slow. Have you tried to run the tests with cache disabled and perhaps even physically remo […]
Show full quote

Ouch. That poor thing is rediculously slow. Have you tried to run the tests with cache disabled and perhaps even physically removed from the board?
I'd parity active or disabled?

For comparison:
I've got a Biostar board with DX40 soldered.
Cachechk: within cache: 37.5 ms/b; outside cache: 78.7 ms/b
Landmark shows 70MHz AT.
My 286-16 even shows up as 24MHz AT
My Asus 486SX with both L1 and L2 disabled shows 14MHz AT in Landmark and 88ms/b in Cachechk.

I'd look into the cache with these figures.

So far disabled cache in bios, then physically remove and finally put back the original 40mhz crystal. Landmark shows the same performance or worse for the 40mhz crystals case and cache check is unchanged.
Thanks

The cache is already disabled. Turbo is ON, On actually means OFF means slow. Turbo in old computers was often controlled by keyboard controller, keyboard controller is right next to battery, leaked battery can damage keyboard controller pins, including one responsible for turbo.

If this the same board you posted about here: Cystal clock 40mhz 386 and similar to https://theretroweb.com/motherboards/s/copam-c3404 ???
then the turbo switch header JP4 may need be jumpered if you already did not. Some boards are reverse, where jumpered is non turbo....but am fairly sure you need to jump it.
If it were me would remove the Math-Co until I got the board running like it should, specially since Landmark says the FPU is running at 1Mhz ....but thats just me 😀

Ok. After jumpering the turbo the landmark test is back at 40mhz.
Cache chips now return 59 u/s but i think i fried the FPU as it reads 4.5mhz.
Bummer

It's possible that the chipset can't handle the higher frequency, but I'd expect different problems if that were the case.